Epidemiology often plays the role of producing data to push others to act. Epidemiological findings may call on medical researchers to identify the mechanism of widespread illness or on policymakers to act to limit risks to the population. The path to change can be a winding one.
Not so with a recent set of papers published by researchers in the UCSF Department of Epidemiology and Biostatistics (DEB) using California death records to understand mortality patterns during the pandemic. In the recent Supreme Court case in which the court was asked to strike down the federal vaccine-or-test mandate for large employers, an American Public Health Association brief cited two DEB studies in its support of the mandate.
The case pitted the Occupational Safety and Health Administration (OSHA) against a business group claiming that OSHA lacked the authority to address COVID because the virus is not specifically an occupational hazard. People might catch COVID at work, the argument went, but they are just as likely to catch it at home or out with friends.
The amicus brief cited one study led by Yea-Hung Chen, PhD, showing that the risk of death during the COVID pandemic varied significantly by occupation. The research, published in PLOS One, found that workers in the food and agriculture industries saw the highest jump in death rates in 2020 compared to what would be expected based on previous years’ data. Workers in transportation and logistics and manufacturing and facilities also saw big jumps in death rates. Specific high-risk jobs included cooks, butchers, couriers and sewing machine operators.
That some industries had significantly bigger upticks in death rates than others led the research team to conclude: “In-person essential work is a likely venue of transmission of coronavirus infection and must be addressed through vaccination and strict enforcement of health orders in workplace settings.”
The researchers considered whether risk within an industry varied by race/ethnicity, finding that it did. Within a given industry, white and Asian workers usually had the lowest mortality while Black workers usually had the highest. This study could not address whether these differences resulted from some groups clustering in certain industries or from a racialized hierarchy of job roles within industries.
A second study cited in the brief – this one led by postdoctoral fellow Ellicott Matthay, PhD, and posted as a pre-print – focused on the extent to which racial disparities in COVID-19 deaths were explained by differences in education and/or access to safer workplaces. The study used death certificates along with a 2019 Census update to identify the race, gender, occupation and educational attainment of everyone living in the state – and thus at risk of dying – during the first phase of the pandemic.
“It was clear from Yea-Hung’s paper that occupation was an important factor in racial disparities, but it was really hard to actually quantify that role,” Matthay explained.
To parse how much of COVID’s racially disparate impact would remain if work was not the driving factor, Matthay and her colleagues considered two hypothetical situations. To measure the role that riskier jobs played driving higher death rates, the team statistically replaced the jobs held by members of non-white groups with those held by white people, who had the lowest risk of COVID death. Then, to consider the overall role work played in driving the racial/ethnic disparities in COVID deaths, the researchers considered how many people would have died if all working-age Californians, regardless of race/ethnicity, had been able to telework earning high pay.
Eliminating the educational and workplace differences between racial groups proved not to have the kind effects one would expect. In California, Asians have higher average educational attainment than whites, so matching them to whites would have made their outcomes worse. Black men work in more dangerous industries, but they also have significantly higher rates of unemployment, so matching their employment to white men’s would have made almost no difference. Black women and Latinx men and women would, however, have decreased their COVID-related risk of death fall by 8-23%, the experiment showed.
Education and telework had the most dramatic impact on people’s risk, the study revealed.
By calculating risk as if all workers were in higher-wage, higher-education, non-essential telework-able and jobs cut working-age deaths by more than half.
“If we afforded the same protection that we do to people with social advantage to the most vulnerable workers, then we could have prevented a lot of deaths,” Matthay said.
The Supreme Court, unfortunately, decided not to increase protection in higher-risk occupations, finding instead that COVID is not an occupational risk that OSHA can regulate. But these studies are still being considered in other policy conversations, Chen and Matthay said.